Face Generation¶

In this project, you'll use generative adversarial networks to generate new images of faces.

Get the Data¶

You'll be using two datasets in this project:

MNIST
CelebA

Since the celebA dataset is complex and you're doing GANs in a project for the first time, we want you to test your neural network on MNIST before CelebA. Running the GANs on MNIST will allow you to see how well your model trains sooner.

If you're using FloydHub, set data_dir to "/input" and use the FloydHub data ID "R5KrjnANiKVhLWAkpXhNBe".

data_dir = './data'

# FloydHub - Use with data ID "R5KrjnANiKVhLWAkpXhNBe"
data_dir = '/input'


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import helper

helper.download_extract('mnist', data_dir)
helper.download_extract('celeba', data_dir)

Found mnist Data
Found celeba Data

Explore the Data¶

MNIST¶

As you're aware, the MNIST dataset contains images of handwritten digits. You can view the first number of examples by changing show_n_images.

show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
%matplotlib inline
import os
from glob import glob
from matplotlib import pyplot

mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'mnist/*.jpg'))[:show_n_images], 28, 28, 'L')
pyplot.imshow(helper.images_square_grid(mnist_images, 'L'), cmap='gray')

<matplotlib.image.AxesImage at 0x7fb1ea3f8048>

CelebA¶

The CelebFaces Attributes Dataset (CelebA) dataset contains over 200,000 celebrity images with annotations. Since you're going to be generating faces, you won't need the annotations. You can view the first number of examples by changing show_n_images.

show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'img_align_celeba/*.jpg'))[:show_n_images], 128, 128, 'RGB')
pyplot.imshow(helper.images_square_grid(mnist_images, 'RGB'))

<matplotlib.image.AxesImage at 0x7fb1ea331a58>

len(glob(os.path.join(data_dir, 'img_align_celeba/*.jpg')))

202599

Preprocess the Data¶

Since the project's main focus is on building the GANs, we'll preprocess the data for you. The values of the MNIST and CelebA dataset will be in the range of -0.5 to 0.5 of 28x28 dimensional images. The CelebA images will be cropped to remove parts of the image that don't include a face, then resized down to 28x28.

The MNIST images are black and white images with a single color channel while the CelebA images have 3 color channels (RGB color channel).

Build the Neural Network¶

You'll build the components necessary to build a GANs by implementing the following functions below:

model_inputs
discriminator
generator
model_loss
model_opt
train

Check the Version of TensorFlow and Access to GPU¶

This will check to make sure you have the correct version of TensorFlow and access to a GPU

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
from distutils.version import LooseVersion
import warnings
import tensorflow as tf

# Check TensorFlow Version
assert LooseVersion(tf.__version__) >= LooseVersion('1.0'), 'Please use TensorFlow version 1.0 or newer.  You are using {}'.format(tf.__version__)
print('TensorFlow Version: {}'.format(tf.__version__))

# Check for a GPU
if not tf.test.gpu_device_name():
    warnings.warn('No GPU found. Please use a GPU to train your neural network.')
else:
    print('Default GPU Device: {}'.format(tf.test.gpu_device_name()))

TensorFlow Version: 1.0.0
Default GPU Device: /gpu:0

Input¶

Implement the model_inputs function to create TF Placeholders for the Neural Network. It should create the following placeholders:

Real input images placeholder with rank 4 using image_width, image_height, and image_channels.
Z input placeholder with rank 2 using z_dim.
Learning rate placeholder with rank 0.

Return the placeholders in the following the tuple (tensor of real input images, tensor of z data)

import problem_unittests as tests

def model_inputs(image_width, image_height, image_channels, z_dim):
    """
    Create the model inputs
    :param image_width: The input image width
    :param image_height: The input image height
    :param image_channels: The number of image channels
    :param z_dim: The dimension of Z
    :return: Tuple of (tensor of real input images, tensor of z data, learning rate)
    """
    # TODO: Implement Function
    input_images = tf.placeholder(tf.float32, [None, image_width, image_height, image_channels], name='input_images')
    z = tf.placeholder(tf.float32, [None, z_dim], name='z')
    learning_rate = tf.placeholder(tf.float32, name='learning_rate')

    return input_images, z, learning_rate


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_inputs(model_inputs)

Tests Passed

Discriminator¶

Implement discriminator to create a discriminator neural network that discriminates on images. This function should be able to reuse the variabes in the neural network. Use tf.variable_scope with a scope name of "discriminator" to allow the variables to be reused. The function should return a tuple of (tensor output of the generator, tensor logits of the generator).

def leaky_relu(x, alpha=0.2, name='leaky_relu'):
            return tf.maximum(x, alpha * x, name=name)

def discriminator(images, reuse=False):
    """
    Create the discriminator network
    :param image: Tensor of input image(s)
    :param reuse: Boolean if the weights should be reused
    :return: Tuple of (tensor output of the discriminator, tensor logits of the discriminator)
    """
    # TODO: Implement Function
    with tf.variable_scope('discriminator', reuse=reuse): # finish this
        
        # Input layer is 28x28x3
        x1 = tf.layers.conv2d(images, 64, 5, strides=2, padding='same', kernel_initializer=tf.contrib.layers.xavier_initializer())
        relu1 = leaky_relu(x1)
        # 14x14x64
        
        x2 = tf.layers.conv2d(relu1, 128, 5, strides=2, padding='same', kernel_initializer=tf.contrib.layers.xavier_initializer())
        bn2 = tf.layers.batch_normalization(x2, training=True)
        relu2 = leaky_relu(bn2)
        # 7x7x128
        
        x3 = tf.layers.conv2d(relu2, 256, 5, strides=2, padding='same', kernel_initializer=tf.contrib.layers.xavier_initializer())
        bn3 = tf.layers.batch_normalization(x3, training=True)
        relu3 = leaky_relu(bn3)
        relu3 = tf.nn.dropout(relu3, 0.9)
        # 4x4x256

        # Flatten it
        flat = tf.reshape(relu3, (-1, 4*4*256))
        logits = tf.layers.dense(flat, 1)
        out = tf.sigmoid(logits)        
        
        
        return out, logits


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_discriminator(discriminator, tf)

Tests Passed

Generator¶

Implement generator to generate an image using z. This function should be able to reuse the variabes in the neural network. Use tf.variable_scope with a scope name of "generator" to allow the variables to be reused. The function should return the generated 28 x 28 x out_channel_dim images.

def generator(z, out_channel_dim, is_train=True):
    """
    Create the generator network
    :param z: Input z
    :param out_channel_dim: The number of channels in the output image
    :param is_train: Boolean if generator is being used for training
    :return: The tensor output of the generator
    """
    # TODO: Implement Function
    
    with tf.variable_scope('generator', reuse=(not is_train)):
        
        # First fully connected layer
        x1 = tf.layers.dense(z, 7*7*512)

        # Reshape it to start the convolutional stack
        x1 = tf.reshape(x1, (-1, 7, 7, 512))
        x1 = tf.layers.batch_normalization(x1, training=is_train)
        x1 = leaky_relu(x1)
        x1 = tf.nn.dropout(x1, 0.5)
        # 7x7x512 now
        
        x2 = tf.layers.conv2d_transpose(x1, 256, 5, strides=1, padding='same')
        x2 = tf.layers.batch_normalization(x2, training=is_train)
        x2 = leaky_relu(x2)
        #x2 = tf.nn.dropout(x2, 0.5)
        # 7x7x256 now
        
        x3 = tf.layers.conv2d_transpose(x2, 128, 5, strides=2, padding='same')
        x3 = tf.layers.batch_normalization(x3, training=is_train)
        x3 = leaky_relu(x3)
        x3 = tf.nn.dropout(x3, 0.5)
        # 14x14x128 now
        
        # Output layer
        logits = tf.layers.conv2d_transpose(x3, out_channel_dim, 5, strides=2, padding='same')
        # 28x28x3 now
        
        out = tf.tanh(logits)
        
        return out

"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_generator(generator, tf)

Tests Passed

Loss¶

Implement model_loss to build the GANs for training and calculate the loss. The function should return a tuple of (discriminator loss, generator loss). Use the following functions you implemented:

discriminator(images, reuse=False)
generator(z, out_channel_dim, is_train=True)

def model_loss(input_real, input_z, out_channel_dim):
    """
    Get the loss for the discriminator and generator
    :param input_real: Images from the real dataset
    :param input_z: Z input
    :param out_channel_dim: The number of channels in the output image
    :return: A tuple of (discriminator loss, generator loss)
    """
    # TODO: Implement Function
    g_model = generator(input_z, out_channel_dim, is_train=True)
    d_model_real, d_logits_real = discriminator(input_real)
    d_model_fake, d_logits_fake = discriminator(g_model, reuse=True)

    d_loss_real = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_real, labels=tf.ones_like(d_model_real) * 0.9))
    d_loss_fake = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, labels=tf.zeros_like(d_model_fake) * 0.9))
    g_loss = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, labels=tf.ones_like(d_model_fake) * 0.9))

    d_loss = d_loss_real + d_loss_fake

    return d_loss, g_loss

"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_loss(model_loss)

Tests Passed

Optimization¶

Implement model_opt to create the optimization operations for the GANs. Use tf.trainable_variables to get all the trainable variables. Filter the variables with names that are in the discriminator and generator scope names. The function should return a tuple of (discriminator training operation, generator training operation).

def model_opt(d_loss, g_loss, learning_rate, beta1):
    """
    Get optimization operations
    :param d_loss: Discriminator loss Tensor
    :param g_loss: Generator loss Tensor
    :param learning_rate: Learning Rate Placeholder
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :return: A tuple of (discriminator training operation, generator training operation)
    """
    # TODO: Implement Function
    # Get weights and bias to update
    t_vars = tf.trainable_variables()
    d_vars = [var for var in t_vars if var.name.startswith('discriminator')]
    g_vars = [var for var in t_vars if var.name.startswith('generator')]

    # Optimize
    with tf.control_dependencies(tf.get_collection(tf.GraphKeys.UPDATE_OPS)):
        d_train_opt = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(d_loss, var_list=d_vars)
        g_train_opt = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(g_loss, var_list=g_vars)

    return d_train_opt, g_train_opt

"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_opt(model_opt, tf)

Tests Passed

Neural Network Training¶

Show Output¶

Use this function to show the current output of the generator during training. It will help you determine how well the GANs is training.

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import numpy as np

def show_generator_output(sess, n_images, input_z, out_channel_dim, image_mode):
    """
    Show example output for the generator
    :param sess: TensorFlow session
    :param n_images: Number of Images to display
    :param input_z: Input Z Tensor
    :param out_channel_dim: The number of channels in the output image
    :param image_mode: The mode to use for images ("RGB" or "L")
    """
    cmap = None if image_mode == 'RGB' else 'gray'
    z_dim = input_z.get_shape().as_list()[-1]
    example_z = np.random.uniform(-1, 1, size=[n_images, z_dim])

    samples = sess.run(
        generator(input_z, out_channel_dim, False),
        feed_dict={input_z: example_z})

    images_grid = helper.images_square_grid(samples, image_mode)
    pyplot.imshow(images_grid, cmap=cmap)
    pyplot.show()

Train¶

Implement train to build and train the GANs. Use the following functions you implemented:

model_inputs(image_width, image_height, image_channels, z_dim)
model_loss(input_real, input_z, out_channel_dim)
model_opt(d_loss, g_loss, learning_rate, beta1)

Use the show_generator_output to show generator output while you train. Running show_generator_output for every batch will drastically increase training time and increase the size of the notebook. It's recommended to print the generator output every 100 batches.

def train(epoch_count, batch_size, z_dim, learning_rate, beta1, get_batches, data_shape, data_image_mode):
    """
    Train the GAN
    :param epoch_count: Number of epochs
    :param batch_size: Batch Size
    :param z_dim: Z dimension
    :param learning_rate: Learning Rate
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :param get_batches: Function to get batches
    :param data_shape: Shape of the data
    :param data_image_mode: The image mode to use for images ("RGB" or "L")
    """
    # TODO: Build Model
    #print(data_shape)
    #print(data_shape[1], data_shape[2])
    #print(data_image_mode)
    #print(z_dim)
    
    tf.reset_default_graph()

    image_channels = 3 if data_image_mode == 'RGB' else 1
    input_real, input_z, tmp = model_inputs(data_shape[1], data_shape[2], image_channels, z_dim)
    d_loss, g_loss = model_loss(input_real, input_z, image_channels)
    d_opt, g_opt = model_opt(d_loss, g_loss, learning_rate, beta1)    

    
    sample_z = np.random.uniform(-1, 1, size=(72, z_dim))
    samples, losses = [], []
    steps = 0
    
    with tf.Session() as sess:
        sess.run(tf.global_variables_initializer())
        for epoch_i in range(epoch_count):
            for batch_images in get_batches(batch_size):
                # TODO: Train Model
                steps += 1
                
                #Normalize real images
                batch_images = batch_images * 2
                
                # Sample random noise for G
                batch_z = np.random.uniform(-1, 1, size=(batch_size, z_dim))

                # Run optimizers
                _ = sess.run(d_opt, feed_dict={input_real: batch_images, input_z: batch_z})
                _ = sess.run(g_opt, feed_dict={input_z: batch_z, input_real: batch_images})

                if steps % print_every == 0:
                    # At the end of each epoch, get the losses and print them out
                    train_loss_d = d_loss.eval({input_z: batch_z, input_real: batch_images})
                    train_loss_g = g_loss.eval({input_z: batch_z})

                    print("Epoch {}/{}...".format(epoch_i+1, epoch_count),
                          "Discriminator Loss: {:.4f}...".format(train_loss_d),
                          "Generator Loss: {:.4f}".format(train_loss_g))
                    # Save losses to view after training
                    losses.append((train_loss_d, train_loss_g))

                if steps % show_every == 0:
                    show_generator_output(sess, n_images, input_z, image_channels, data_image_mode)

MNIST¶

Test your GANs architecture on MNIST. After 2 epochs, the GANs should be able to generate images that look like handwritten digits. Make sure the loss of the generator is lower than the loss of the discriminator or close to 0.

batch_size = 64
z_dim = 100
learning_rate = 0.0002
beta1 = 0.5

print_every=10
show_every=100
n_images=72

"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 2

mnist_dataset = helper.Dataset('mnist', glob(os.path.join(data_dir, 'mnist/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, mnist_dataset.get_batches,
          mnist_dataset.shape, mnist_dataset.image_mode)

Epoch 1/2... Discriminator Loss: 0.6245... Generator Loss: 2.9767
Epoch 1/2... Discriminator Loss: 1.0137... Generator Loss: 0.9982
Epoch 1/2... Discriminator Loss: 1.7816... Generator Loss: 0.5153
Epoch 1/2... Discriminator Loss: 0.6495... Generator Loss: 1.5672
Epoch 1/2... Discriminator Loss: 0.4990... Generator Loss: 2.6239
Epoch 1/2... Discriminator Loss: 0.5766... Generator Loss: 1.9216
Epoch 1/2... Discriminator Loss: 0.8263... Generator Loss: 3.4052
Epoch 1/2... Discriminator Loss: 0.9474... Generator Loss: 1.0935
Epoch 1/2... Discriminator Loss: 0.9372... Generator Loss: 1.2252
Epoch 1/2... Discriminator Loss: 0.6996... Generator Loss: 2.1619

Epoch 1/2... Discriminator Loss: 0.6912... Generator Loss: 1.5775
Epoch 1/2... Discriminator Loss: 0.6713... Generator Loss: 2.5430
Epoch 1/2... Discriminator Loss: 0.9356... Generator Loss: 2.7339
Epoch 1/2... Discriminator Loss: 0.9246... Generator Loss: 2.4588
Epoch 1/2... Discriminator Loss: 0.7745... Generator Loss: 2.3736
Epoch 1/2... Discriminator Loss: 0.8916... Generator Loss: 1.5773
Epoch 1/2... Discriminator Loss: 0.8451... Generator Loss: 1.2032
Epoch 1/2... Discriminator Loss: 0.9747... Generator Loss: 0.9371
Epoch 1/2... Discriminator Loss: 0.7633... Generator Loss: 2.0545
Epoch 1/2... Discriminator Loss: 0.8650... Generator Loss: 1.1298

Epoch 1/2... Discriminator Loss: 0.7974... Generator Loss: 1.3735
Epoch 1/2... Discriminator Loss: 0.9134... Generator Loss: 1.1237
Epoch 1/2... Discriminator Loss: 1.2249... Generator Loss: 0.7630
Epoch 1/2... Discriminator Loss: 0.7537... Generator Loss: 1.2057
Epoch 1/2... Discriminator Loss: 0.8802... Generator Loss: 1.4971
Epoch 1/2... Discriminator Loss: 0.7176... Generator Loss: 1.4501
Epoch 1/2... Discriminator Loss: 0.9341... Generator Loss: 2.5643
Epoch 1/2... Discriminator Loss: 1.2729... Generator Loss: 3.1533
Epoch 1/2... Discriminator Loss: 0.7595... Generator Loss: 2.0608
Epoch 1/2... Discriminator Loss: 0.7426... Generator Loss: 1.3723

Epoch 1/2... Discriminator Loss: 0.7222... Generator Loss: 1.6877
Epoch 1/2... Discriminator Loss: 0.9035... Generator Loss: 2.3029
Epoch 1/2... Discriminator Loss: 0.7562... Generator Loss: 1.2797
Epoch 1/2... Discriminator Loss: 0.8669... Generator Loss: 1.0032
Epoch 1/2... Discriminator Loss: 0.6931... Generator Loss: 1.5528
Epoch 1/2... Discriminator Loss: 0.6310... Generator Loss: 1.6585
Epoch 1/2... Discriminator Loss: 0.8270... Generator Loss: 1.1692
Epoch 1/2... Discriminator Loss: 0.7901... Generator Loss: 2.0759
Epoch 1/2... Discriminator Loss: 0.7465... Generator Loss: 1.6035
Epoch 1/2... Discriminator Loss: 1.3024... Generator Loss: 0.6725

Epoch 1/2... Discriminator Loss: 0.6227... Generator Loss: 1.8583
Epoch 1/2... Discriminator Loss: 0.8924... Generator Loss: 2.8623
Epoch 1/2... Discriminator Loss: 1.2388... Generator Loss: 0.6938
Epoch 1/2... Discriminator Loss: 1.0535... Generator Loss: 0.8355
Epoch 1/2... Discriminator Loss: 0.7898... Generator Loss: 2.4401
Epoch 1/2... Discriminator Loss: 1.1715... Generator Loss: 0.8321
Epoch 1/2... Discriminator Loss: 0.9731... Generator Loss: 2.8267
Epoch 1/2... Discriminator Loss: 0.7151... Generator Loss: 1.5583
Epoch 1/2... Discriminator Loss: 0.8225... Generator Loss: 2.1567
Epoch 1/2... Discriminator Loss: 0.6531... Generator Loss: 1.7604

Epoch 1/2... Discriminator Loss: 0.6940... Generator Loss: 1.3514
Epoch 1/2... Discriminator Loss: 0.6743... Generator Loss: 1.4088
Epoch 1/2... Discriminator Loss: 0.7445... Generator Loss: 1.5115
Epoch 1/2... Discriminator Loss: 0.7470... Generator Loss: 1.3462
Epoch 1/2... Discriminator Loss: 0.9697... Generator Loss: 2.1914
Epoch 1/2... Discriminator Loss: 0.7342... Generator Loss: 1.5046
Epoch 1/2... Discriminator Loss: 0.8645... Generator Loss: 2.1115
Epoch 1/2... Discriminator Loss: 0.8176... Generator Loss: 1.1193
Epoch 1/2... Discriminator Loss: 0.8198... Generator Loss: 1.2709
Epoch 1/2... Discriminator Loss: 0.7651... Generator Loss: 1.7586

Epoch 1/2... Discriminator Loss: 0.7812... Generator Loss: 1.4344
Epoch 1/2... Discriminator Loss: 0.9991... Generator Loss: 0.9287
Epoch 1/2... Discriminator Loss: 1.0186... Generator Loss: 2.8117
Epoch 1/2... Discriminator Loss: 0.8363... Generator Loss: 1.0562
Epoch 1/2... Discriminator Loss: 0.8522... Generator Loss: 1.1088
Epoch 1/2... Discriminator Loss: 0.8408... Generator Loss: 1.1924
Epoch 1/2... Discriminator Loss: 0.9341... Generator Loss: 0.9994
Epoch 1/2... Discriminator Loss: 0.7437... Generator Loss: 2.0655
Epoch 1/2... Discriminator Loss: 0.8869... Generator Loss: 1.0442
Epoch 1/2... Discriminator Loss: 1.1579... Generator Loss: 2.4525

Epoch 1/2... Discriminator Loss: 0.7294... Generator Loss: 1.3088
Epoch 1/2... Discriminator Loss: 0.7150... Generator Loss: 2.2776
Epoch 1/2... Discriminator Loss: 0.7130... Generator Loss: 1.3028
Epoch 1/2... Discriminator Loss: 0.8543... Generator Loss: 1.1291
Epoch 1/2... Discriminator Loss: 0.7428... Generator Loss: 1.5942
Epoch 1/2... Discriminator Loss: 0.7935... Generator Loss: 1.3099
Epoch 1/2... Discriminator Loss: 0.7547... Generator Loss: 1.5581
Epoch 1/2... Discriminator Loss: 0.6600... Generator Loss: 1.9245
Epoch 1/2... Discriminator Loss: 1.0393... Generator Loss: 0.8040
Epoch 1/2... Discriminator Loss: 1.1287... Generator Loss: 0.8268

Epoch 1/2... Discriminator Loss: 0.8168... Generator Loss: 2.6359
Epoch 1/2... Discriminator Loss: 0.7622... Generator Loss: 1.8338
Epoch 1/2... Discriminator Loss: 0.8179... Generator Loss: 2.1429
Epoch 1/2... Discriminator Loss: 0.7098... Generator Loss: 1.8643
Epoch 1/2... Discriminator Loss: 0.9634... Generator Loss: 2.5022
Epoch 1/2... Discriminator Loss: 0.8073... Generator Loss: 1.5668
Epoch 1/2... Discriminator Loss: 1.1385... Generator Loss: 1.0048
Epoch 1/2... Discriminator Loss: 0.7744... Generator Loss: 1.8734
Epoch 1/2... Discriminator Loss: 0.7984... Generator Loss: 1.7751
Epoch 1/2... Discriminator Loss: 0.7908... Generator Loss: 2.1333

Epoch 1/2... Discriminator Loss: 0.7909... Generator Loss: 2.1846
Epoch 1/2... Discriminator Loss: 0.7712... Generator Loss: 1.3055
Epoch 1/2... Discriminator Loss: 0.7558... Generator Loss: 1.5152
Epoch 2/2... Discriminator Loss: 0.7651... Generator Loss: 1.0291
Epoch 2/2... Discriminator Loss: 0.7049... Generator Loss: 1.4845
Epoch 2/2... Discriminator Loss: 0.6775... Generator Loss: 1.3491
Epoch 2/2... Discriminator Loss: 0.8421... Generator Loss: 1.0910
Epoch 2/2... Discriminator Loss: 0.5960... Generator Loss: 2.3230
Epoch 2/2... Discriminator Loss: 0.9060... Generator Loss: 1.1126
Epoch 2/2... Discriminator Loss: 0.7660... Generator Loss: 1.3537

CelebA¶

Run your GANs on CelebA. It will take around 20 minutes on the average GPU to run one epoch. You can run the whole epoch or stop when it starts to generate realistic faces.

batch_size = 64
z_dim = 100
learning_rate = 0.0002
beta1 = 0.5

print_every=1
show_every=10
n_images=72

"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 1

celeba_dataset = helper.Dataset('celeba', glob(os.path.join(data_dir, 'img_align_celeba/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, celeba_dataset.get_batches,
          celeba_dataset.shape, celeba_dataset.image_mode)

Epoch 1/1... Discriminator Loss: 1.6370... Generator Loss: 0.6965
Epoch 1/1... Discriminator Loss: 0.9613... Generator Loss: 1.1012
Epoch 1/1... Discriminator Loss: 0.8910... Generator Loss: 1.4954
Epoch 1/1... Discriminator Loss: 0.6458... Generator Loss: 1.7962
Epoch 1/1... Discriminator Loss: 0.6594... Generator Loss: 1.7604
Epoch 1/1... Discriminator Loss: 0.5916... Generator Loss: 2.0987
Epoch 1/1... Discriminator Loss: 0.5110... Generator Loss: 2.5220
Epoch 1/1... Discriminator Loss: 0.5584... Generator Loss: 2.0564
Epoch 1/1... Discriminator Loss: 0.5264... Generator Loss: 2.2422
Epoch 1/1... Discriminator Loss: 0.7018... Generator Loss: 1.6343

Epoch 1/1... Discriminator Loss: 0.4708... Generator Loss: 3.2848
Epoch 1/1... Discriminator Loss: 0.5119... Generator Loss: 2.7142
Epoch 1/1... Discriminator Loss: 0.6466... Generator Loss: 1.6043
Epoch 1/1... Discriminator Loss: 0.5045... Generator Loss: 3.6357
Epoch 1/1... Discriminator Loss: 0.4766... Generator Loss: 2.7198
Epoch 1/1... Discriminator Loss: 0.5641... Generator Loss: 2.2540
Epoch 1/1... Discriminator Loss: 0.4674... Generator Loss: 3.1834
Epoch 1/1... Discriminator Loss: 0.6882... Generator Loss: 1.5283
Epoch 1/1... Discriminator Loss: 0.4084... Generator Loss: 4.0102
Epoch 1/1... Discriminator Loss: 0.4152... Generator Loss: 3.3829

Epoch 1/1... Discriminator Loss: 0.5841... Generator Loss: 1.7559
Epoch 1/1... Discriminator Loss: 0.5512... Generator Loss: 4.3501
Epoch 1/1... Discriminator Loss: 0.5637... Generator Loss: 2.1702
Epoch 1/1... Discriminator Loss: 0.4291... Generator Loss: 3.2059
Epoch 1/1... Discriminator Loss: 0.4851... Generator Loss: 2.5983
Epoch 1/1... Discriminator Loss: 0.5854... Generator Loss: 1.7212
Epoch 1/1... Discriminator Loss: 0.4383... Generator Loss: 4.4280
Epoch 1/1... Discriminator Loss: 0.4317... Generator Loss: 3.2620
Epoch 1/1... Discriminator Loss: 0.4403... Generator Loss: 2.9409
Epoch 1/1... Discriminator Loss: 0.4330... Generator Loss: 3.4628

Epoch 1/1... Discriminator Loss: 0.4538... Generator Loss: 2.6600
Epoch 1/1... Discriminator Loss: 0.4190... Generator Loss: 3.7027
Epoch 1/1... Discriminator Loss: 0.4123... Generator Loss: 3.0996
Epoch 1/1... Discriminator Loss: 0.4116... Generator Loss: 3.2595
Epoch 1/1... Discriminator Loss: 0.4027... Generator Loss: 3.4308
Epoch 1/1... Discriminator Loss: 0.8103... Generator Loss: 1.2528
Epoch 1/1... Discriminator Loss: 0.6628... Generator Loss: 7.2769
Epoch 1/1... Discriminator Loss: 0.4528... Generator Loss: 3.7128
Epoch 1/1... Discriminator Loss: 0.9323... Generator Loss: 1.1248
Epoch 1/1... Discriminator Loss: 1.0316... Generator Loss: 7.0197

Epoch 1/1... Discriminator Loss: 0.5436... Generator Loss: 2.4296
Epoch 1/1... Discriminator Loss: 0.7828... Generator Loss: 1.5364
Epoch 1/1... Discriminator Loss: 0.5224... Generator Loss: 5.9282
Epoch 1/1... Discriminator Loss: 0.4034... Generator Loss: 4.7734
Epoch 1/1... Discriminator Loss: 0.5879... Generator Loss: 1.9454
Epoch 1/1... Discriminator Loss: 0.4642... Generator Loss: 3.4416
Epoch 1/1... Discriminator Loss: 0.4422... Generator Loss: 3.2819
Epoch 1/1... Discriminator Loss: 0.4602... Generator Loss: 2.6964
Epoch 1/1... Discriminator Loss: 0.4313... Generator Loss: 3.4865
Epoch 1/1... Discriminator Loss: 0.4300... Generator Loss: 3.3255

Epoch 1/1... Discriminator Loss: 0.5736... Generator Loss: 1.6873
Epoch 1/1... Discriminator Loss: 0.6000... Generator Loss: 4.8824
Epoch 1/1... Discriminator Loss: 0.4875... Generator Loss: 2.3594
Epoch 1/1... Discriminator Loss: 0.4982... Generator Loss: 2.2208
Epoch 1/1... Discriminator Loss: 0.5556... Generator Loss: 4.9266
Epoch 1/1... Discriminator Loss: 0.7630... Generator Loss: 1.2790
Epoch 1/1... Discriminator Loss: 0.6995... Generator Loss: 6.8085
Epoch 1/1... Discriminator Loss: 0.4526... Generator Loss: 3.4383
Epoch 1/1... Discriminator Loss: 0.9558... Generator Loss: 1.0886
Epoch 1/1... Discriminator Loss: 1.2195... Generator Loss: 7.5255

Epoch 1/1... Discriminator Loss: 0.5936... Generator Loss: 2.3203
Epoch 1/1... Discriminator Loss: 0.8476... Generator Loss: 1.1442
Epoch 1/1... Discriminator Loss: 0.7777... Generator Loss: 7.7971
Epoch 1/1... Discriminator Loss: 0.5516... Generator Loss: 5.3153
Epoch 1/1... Discriminator Loss: 0.7325... Generator Loss: 1.5436
Epoch 1/1... Discriminator Loss: 0.5006... Generator Loss: 3.0562
Epoch 1/1... Discriminator Loss: 0.5180... Generator Loss: 3.5800
Epoch 1/1... Discriminator Loss: 0.4683... Generator Loss: 2.8494
Epoch 1/1... Discriminator Loss: 0.5484... Generator Loss: 1.9590
Epoch 1/1... Discriminator Loss: 0.4355... Generator Loss: 3.7666

Epoch 1/1... Discriminator Loss: 0.4778... Generator Loss: 2.3629
Epoch 1/1... Discriminator Loss: 0.4529... Generator Loss: 2.5842
Epoch 1/1... Discriminator Loss: 0.4278... Generator Loss: 3.3964
Epoch 1/1... Discriminator Loss: 0.4410... Generator Loss: 3.2068
Epoch 1/1... Discriminator Loss: 0.4419... Generator Loss: 2.6210
Epoch 1/1... Discriminator Loss: 0.4558... Generator Loss: 2.7147
Epoch 1/1... Discriminator Loss: 0.6077... Generator Loss: 1.6903
Epoch 1/1... Discriminator Loss: 0.4638... Generator Loss: 3.0644
Epoch 1/1... Discriminator Loss: 0.4539... Generator Loss: 2.6861
Epoch 1/1... Discriminator Loss: 0.4655... Generator Loss: 2.1721

Epoch 1/1... Discriminator Loss: 0.4456... Generator Loss: 2.8869
Epoch 1/1... Discriminator Loss: 0.6241... Generator Loss: 1.5674
Epoch 1/1... Discriminator Loss: 0.5765... Generator Loss: 2.8183
Epoch 1/1... Discriminator Loss: 0.5921... Generator Loss: 1.6445
Epoch 1/1... Discriminator Loss: 0.6389... Generator Loss: 1.7947
Epoch 1/1... Discriminator Loss: 0.5820... Generator Loss: 2.4041
Epoch 1/1... Discriminator Loss: 0.6177... Generator Loss: 1.5753
Epoch 1/1... Discriminator Loss: 0.7748... Generator Loss: 3.9352
Epoch 1/1... Discriminator Loss: 0.9124... Generator Loss: 1.0979
Epoch 1/1... Discriminator Loss: 1.0862... Generator Loss: 6.1764

Epoch 1/1... Discriminator Loss: 0.9815... Generator Loss: 1.9027
Epoch 1/1... Discriminator Loss: 0.6059... Generator Loss: 2.6682
Epoch 1/1... Discriminator Loss: 0.6602... Generator Loss: 3.4870
Epoch 1/1... Discriminator Loss: 0.7282... Generator Loss: 1.8330
Epoch 1/1... Discriminator Loss: 0.5781... Generator Loss: 2.7422
Epoch 1/1... Discriminator Loss: 0.6032... Generator Loss: 1.7950
Epoch 1/1... Discriminator Loss: 0.5502... Generator Loss: 2.5282
Epoch 1/1... Discriminator Loss: 0.5302... Generator Loss: 2.0189
Epoch 1/1... Discriminator Loss: 0.5224... Generator Loss: 3.1828
Epoch 1/1... Discriminator Loss: 0.4956... Generator Loss: 2.5254

Submitting This Project¶

When submitting this project, make sure to run all the cells before saving the notebook. Save the notebook file as "dlnd_face_generation.ipynb" and save it as a HTML file under "File" -> "Download as". Include the "helper.py" and "problem_unittests.py" files in your submission.